Gemini 3.1 Flash Lite Image and Claude Sonnet 5 are now on CometAPI — fast, cost-effective image generation and editing, plus high-performance coding and agent workflows. Try them now
GLM 5.2
Z

GLM 5.2

入力:$1.12/M
出力:$3.528/M
GLM-5.2は、オープンソースの大規模モデルとAIコーディングの分野におけるZhipuによる重要なアップデートです。
GLM 5.1
Z

GLM 5.1

入力:$1.12/M
出力:$3.528/M
GLM-5.1(2026年4月リリース)は、長期的な自律タスク向けに専用設計されています。短い対話に最適化された従来のモデルとは異なり、GLM-5.1は目標の整合性を維持し、戦略の逸脱を抑制し、長期間にわたってプロダクショングレードの成果を提供します—単一の複雑なタスクに対して最大8時間の連続自律実行が可能です。これはエージェントエンジニアリングにおける大きな飛躍であり、評価を単一ターンの知能から実世界での持続的な実行へとシフトさせます。
GLM 5 Turbo
Z

GLM 5 Turbo

文脈:200k
入力:$0.96/M
出力:$3.264/M
GLM-5 Turbo は、OpenClaw シナリオのようなエージェント駆動型環境において、高速な推論と優れた性能を実現するよう設計された、Z.ai の新しいモデルです。
GLM 5
Z

GLM 5

入力:$0.8/M
出力:$3.2/M
Want this rewritten or shaped for a specific use? Here are quick variants—pick one to iterate. - Tagline: GLM-5: the open-source model that builds and runs complex systems, not just code. - Hero (2 sentences): GLM-5 is Z.ai’s open-source foundation model for expert developers tackling large-scale software. With agentic planning, deep backend reasoning, and iterative self-correction, it assembles full systems and executes long-horizon workflows with production-grade reliability. - Crisp rewrite: GLM-5 is Z.ai’s flagship open-source model for complex system design and long-horizon agents. It rivals leading closed models on large-scale programming, using agentic planning, deep backend reasoning, and self-correction to go beyond code generation to full-system construction and autonomous execution. - Technical datasheet style: GLM-5 is an OSS code+systems model optimized for long-horizon agents: hierarchical planning, tool orchestration, multi-service integration, and runtime control loops. Built for production workloads on large repos and backends; focuses on end-to-end system construction and autonomous execution. - Enterprise angle: Bring closed-model performance in-house. GLM-5 delivers open, auditable long-horizon agents—planning, integration, and autonomous execution across services—designed for production scale. - OSS/README: Open-source and builder-first, GLM-5 ships agentic planning, deep reasoning, and self-correction so you can go from prompts to running systems—contributions welcome. If this is for a landing page or datasheet, consider adding concrete proof points: key benchmarks (e.g., SWE-bench, HumanEval pass@1), max context, tool-use/runtime support, latency/TPS on common GPUs, license, API compatibility, and example workflows. Which tone/length should I polish?